The Reason Foundation has a fantastic article on the systemic, behavioral bias at play in making public pension actuarial assumptions.
The article cites a recent study that reveals, contrary to popular opinion, investment returns only play a small part in long term health of the fund. In fact, most plans would still have significant unfunded liabilities, even if they hit their investment goals. How could this be?
Stepping back from all of this for a moment, and looking at board decision-making, leads to the point of this being a behavioral issue at work that results from structural governance issues. Nearly every person at the decision-making level benefits from low projected normal costs, and nobody at the decision-making level is harmed by high unfunded liability amortization payments in the short term.
In fact, there is an incentive to drive down normal costs through the careful selection of actuarial assumptions. This lowers the employee’s contribution rate, effectively increasing take home pay, and, due to established laws and court precedent, irrevocably transfers the entire risk and debt to the government agency – and therefore the taxpayers – through either higher taxes or lower levels of services. There is neither an incentive for being accurate, nor a consequence for being inaccurate. It is actually a perverse system in which there is a win for the entire membership when pension board trustees are wrong!
The whole article is very much worth reading in its entirety.